OpenAI automates security testing with GPT-Red

OpenAI

OpenAI's GPT-Red is an automated red-teaming model that uses self-play reinforcement learning to find AI vulnerabilities at scale, successfully making GPT-5.6 6x more robust to prompt injection attacks. This creates a safety flywheel where today's models help secure tomorrow's systems, addressing the scalability limitations of human red-teaming alone.

TechCrunch

Anthropic and Blackstone launched Ode, a $1.5B AI implementation firm, betting that enterprise AI adoption requires specialized deployment expertise as much as advanced models themselves. This reflects a broader industry shift where AI labs are building separate businesses to help customers actually integrate AI into their operations.

Thinking Machines AI

Thinking Machines released Inkling, a 975B-parameter open-weights MoE model trained on 45 trillion multimodal tokens, designed as a customizable foundation model with controllable thinking effort and strong agentic capabilities. The release includes a smaller 12B-active-parameter variant and is available for fine-tuning on their Tinker platform with broad ecosystem support.

OpenAI

Supply Co. and Work Louder released a $230 specialized keyboard with RGB feedback and physical controls designed specifically for managing AI agent workflows and ChatGPT Codex interactions. The hardware represents emerging demand for purpose-built peripherals as agentic AI becomes embedded in developer workflows.

notes.designarena.ai

GPT-5.6 Sol tops design benchmarks by actively avoiding AI-generated design clichés while balancing templated structures with personalized outputs, offering superior speed and cost efficiency. This marks OpenAI's first leadership position on the Design Arena leaderboard and demonstrates sophisticated learned taste in visual generation.